Ependymoma is a neuroepithelial tumor that occurs predominantly in children and young adults, accounting for 8-10 percent of intracranial tumors in children. Despite recent advances in neurosurgery, radiotherapy and chemotherapy, nearly half of the children with intracranial ependymoma eventually die of progressive disease. This unsatisfactory prognosis reflects our poor understanding of the biology of this tumor. Current prognostic factors such as age at diagnosis and degree of surgical resection are inadequate predictors of outcome. Despite claims that the outcome of differentiated and anaplastic ependymoma are different, the influence of histology on prognosis remains controversial. Thus there is an urgent need to identify reliable prognostic markers to provide a more objective and accurate way to classify ependymomas for treatment stratification in multi-institutional clinical trials. One strategy to discover such prognostic markers is to identify genetic alterations that determine the malignant behavior in tumor cells. We have recently developed new approaches to perform comprehensive genetic analysis on pediatric brain tumors by using state-of-the-art genomic technologies. One technique, comparative genomic hybridization (CGH), enables us to create complete profiles of chromosome copy number aberrations (CNAs) in tumors with high efficiency and precision. Our preliminary results with a small number of cases suggest that these profiles identify clinically relevant subgroups of ependymomas. This proposal will expand our study to provide enough statistical power to draw definitive conclusions. At the same time, we are making major modifications to the standard CGH technique by converting it into a high throughput and more sensitive format. Instead of analyzing the aberrations at the level of chromosomes, we are substituting an array of mapped clones of DNA fragments for individual chromosomes. This latest modification greatly improves the precision and efficiency of CGH, produces a precise physical map of the genetic abnormality and will allow the analysis of more samples in less time. This proposal will be the biologic study arm of a new COG Phase II trial for ependymoma (ACNS-0121) approved by CTEP. Our goal is to build a molecular classification system for ependymomas to allow objective patient stratification for future clinical trials. In addition, identifying the genetic abnormalities in ependymomas may ultimately lead to the discovery of new therapeutic targets.